/srv/irclogs.ubuntu.com/2012/11/29/#maas.txt

bigjoolsroaksoax: around?01:45
shangbigjools: question, for the 12.04-nocobbler, I seen these critical bugs: https://launchpad.net/maas/+milestone/12.04.1-sru. Are these bugs also affect the 12.10 maas version? or only exist in the backported one?03:12
bigjoolsshang: ignore that milestone03:12
bigjoolswe're backporting the https://launchpad.net/maas/+milestone/12.10-stabilization one to precise03:13
shangok03:14
shangso the issues will be fixed in 12.10 first then backport it to 12.04?03:15
bigjoolsshang: yup03:15
shangbigjools: if I read this correctly, the backport hasn't begin yet?03:16
bigjoolsshang: we had some dependencies to backport first, they are in -proposed03:16
bigjoolswhen they are in -updates, we'll put maas in -proposed03:17
shangbigjools: ETA will be January?03:17
bigjoolsbefore the end of this year03:17
shangdo you think these critical issues will be an issue if the customer deploy it in their production environment?03:18
bigjoolsyes03:20
bigjoolsit's why we're backporting all the way back to precise03:21
shangbigjools: I mean the current version of maas in 12.10 for production, since it is not ready in 12.04 yet.03:26
bigjoolsyes, it has problems03:27
shangalso, is there a place where I can track the progress for the backporting?03:27
roaksoaxbigjools: here!03:49
bigjoolsroaksoax: hey!03:49
roaksoaxbigjools: howdy!03:50
roaksoaxhow's everything03:50
bigjoolsroaksoax: just wondering, did you do the apparmor backport to precise?03:50
bigjoolsshang: yes, see https://blueprints.launchpad.net/maas/+spec/cdo-r-maas-12.10-sru03:50
roaksoaxbigjools: apparmor didn;'t need the backport, what needed was isc-dhcp03:50
bigjoolsshang: or the milestone will get marked released when I do it03:50
roaksoaxbigjools: and it was there long time03:50
roaksoaxbigjools: nobody processed it though03:50
bigjoolsroaksoax: oh so already done?03:51
roaksoaxbigjools: yeah it is not released yet03:51
bigjoolsoh, stuck in the queue or in proposed or what?03:51
roaksoaxbigjools: yeah, i can't remember wether its been more than 7 days but i'll make sure it gets done tomorrow03:51
roaksoaxthe only thing left is python-tx-tftp for precise, the rest has been uploaded03:51
roaksoaxnow it is time to do some politics to get things accepted03:52
bigjoolsroaksoax: ok I'll add work items on the blueprint03:52
bigjoolsroaksoax: good luck :)03:52
shangbigjools: thanks03:52
roaksoaxbigjools: https://blueprints.launchpad.net/ubuntu/+spec/servercloud-r-maas-next-steps03:52
roaksoaxbigjools: all the INPROGRESS ones should be the ones in the queue03:52
bigjoolsah ok you have another blueprint03:53
bigjoolsroaksoax: tip top, thanks03:54
roaksoaxbigjools: that blueprint is the public one03:54
roaksoaxbigjools: we usually track things that way03:54
bigjoolsack03:55
roaksoaxbigjools: ok so two major things03:56
roaksoaxbigjools: 1. change how conffiles are being handled03:56
roaksoax2. can't remember :)03:56
bigjoolsyes, conf files are a PITA03:56
bigjoolsa RPITA03:56
roaksoaxbigjools: yeah i'll get that done this week or early next week03:57
bigjoolsroaksoax: the major thing is that 1) we amend them on a reconfigure, 2) we amend them when installing, 3) they always conflict when installing03:58
roaksoaxbigjools: the other thing, is we need to start testing upgrades from 1. older precise -> newer precise 2. older precise -> newer quantal 3. newer precise -> newer quantal 4. older quantal -> raring 5. newer quantal -> raring03:58
bigjoolsI am doing #103:58
roaksoaxbigjools: the approach i'll take is simple. Install them in /usr/share/maas/conf, cp them into /etc/maas in postinst, and do configuration03:59
roaksoaxbigjools: in preinst, it will backup the previous config03:59
bigjoolsthe problem is when changes really do need to be made :/03:59
roaksoaxbigjools: yes, user changes, that's indeed the problem03:59
roaksoaxbut i can't see any other easy way around it, unles we get .d support04:00
roaksoaxfor *all* the configs04:00
roaksoaxcause that should actually be the correct approach of doing things. .d for package modifications04:01
roaksoaxbigjools: .d/ should be simple enough for maas_local_settings and the celery configs, not for the .yaml configs04:04
bigjoolsroaksoax: yeah, the yaml is the main source of pain04:05
roaksoaxif we add .d/ for the other stuff, we can simply treat the yaml configs as the process specified above04:05
bigjoolsit would be worth enumerating the pros and cons for that04:06
roaksoaxalright04:08
bigjoolsroaksoax: because we've made some critical changes to the yaml file in the past and if we had a separate copy installed that didn't get updated, the service would silently fail04:10
roaksoaxbigjools: right, but that's why you have to make sure that the packaging does update04:11
bigjoolsroaksoax: pick your poison :)04:12
roaksoaxbigjools: another thing is that we could use debconf more extensively to store values for re-use on upgrades04:12
bigjoolsmight be easier, dunno04:13
roaksoaxyeah anyways, i don't want to/won't stress anbout it04:13
bigjoolsyeah04:16
roaksoaxanyways i'm off04:18
roaksoaxhave a good day04:18
bigjoolscheerio04:55
AskUbuntuHow to debug maasserver with django runserver | http://askubuntu.com/q/22333007:15
=== mgz_ is now known as mgz
=== matsubara is now known as matsubara-lunch
=== matsubara-lunch is now known as matsubara
AskUbuntuCreating block devices for openstack deployment using MAAS and juju (nova-volume deployment) | http://askubuntu.com/q/22351017:29
=== Guest46508 is now known as dpb___
adam_gis it possible to set kernel parameters for nodes?23:39

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!